Skip to content

Pull requests: google/XNNPACK

New pull request New
208 Open 9,409 Closed
208 Open 9,409 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[RVV] add rvv f32 kernels for vcos,vexp,vlog,vsigmoid,vsin,vtanh
#9926 opened Apr 9, 2026 by ken-unger Loading…
3
Add Neon f32-bf16-vcvt kernels
#9922 opened Apr 8, 2026 by GregoryComer • Draft
This CL introduces several enhancements to the dot product scheduling and parallelism logic:
#9916 opened Apr 8, 2026 by copybara-service bot Loading…
Optimize dot product scheduling and parallelism.
#9910 opened Apr 8, 2026 by copybara-service bot Loading…
math.h for RISCV scalar fp16 respect XNN_ENABLE_RISCV_FP16_VECTOR
#9904 opened Apr 7, 2026 by copybara-service bot Loading…
1
Fix integer truncation in deconvolution kernel_size computation
#9876 opened Apr 4, 2026 by mohammadmseet-hue Loading…
Fix integer overflow in allocation size calculations
#9860 opened Apr 3, 2026 by mohammadmseet-hue Loading…
Reduce windows GH actions runners from 32-core to generic windows-2022 machines
#9854 opened Apr 1, 2026 by copybara-service bot Loading…
Fix stack buffer overflows in NCHW reduce rewrite and ynnpack shim functions
#9851 opened Apr 1, 2026 by mohammadmseet-hue Loading…
1
Fix missing bounds checks in subgraph API functions
#9845 opened Apr 1, 2026 by mohammadmseet-hue Loading…
2 tasks done
Fix stack buffer overflows in ynnpack channelwise quantized tensor and reduce
#9843 opened Apr 1, 2026 by mohammadmseet-hue Loading…
4
Add CMake build for YNNPACK
#9803 opened Mar 27, 2026 by copybara-service bot Loading…
[gn] Fix component builds
#9785 opened Mar 26, 2026 by copybara-service bot Loading…
Add 2 bit SSE GEMM microkernels
#9752 opened Mar 23, 2026 by copybara-service bot Loading…
Add SIMD mask and select operations.
#9749 opened Mar 20, 2026 by copybara-service bot Loading…
VNNI microkernels disable XNN_FORCE_REALIZATION
#9719 opened Mar 18, 2026 by copybara-service bot Loading…
Replicate the scalar scaling factor for 4-bit tensor-wise quantized weights to the number of channels, as is done for 8-bit weights.
#9718 opened Mar 18, 2026 by copybara-service bot Loading…
Added deprecation warnings for fp16 flags in XNNPACK.
#9695 opened Mar 16, 2026 by copybara-service bot Loading…
Add a graph rewrite to fallback to fp32 when fp16 isn't supported.
#9686 opened Mar 13, 2026 by copybara-service bot Loading…
Added non trivial batch dim.
#9684 opened Mar 13, 2026 by copybara-service bot Loading…
[WIP] NEONDOT qd8_bf16_qc8w GEMM microkernels
#9673 opened Mar 12, 2026 by copybara-service bot Loading…
Replaced XNNPACK fp16 rewrite with delegate side fp16 rewrite.
#9669 opened Mar 11, 2026 by copybara-service bot Loading…
[RVV] Add missing maxpool and avgpool rvv kernels
#9622 opened Mar 3, 2026 by ken-unger Loading…
3
[gn] Experimental CI for Windows
#9605 opened Feb 27, 2026 by copybara-service bot Loading…
Fix BF16 dot emulation for dynamic B
#9595 opened Feb 25, 2026 by copybara-service bot Loading…
ProTip! Updated in the last three days: updated:>2026-04-06.